Reseach On Algorithm For Automatic Analysis And Detection Of Abnormal ECG Waveform

Sudden cardiac death(SCD) is a major public health problem and one of leading cause of mortality in the whole word, 80% of which are leaded by ventricular fibrillation(VF) and ventricular tachycardia(VT). The implementation of electric defibrillation(ED) in time is widely used in the clinical for patients of the SCD. These patients often die because they fail to obtain treatments in time. For non-fatal arrhythmias, the implementation of remote ECG monitoring, early detection of patients with abnormal precursor and timely treatment can reduce mortality and morbidity by nearly half. The emergence of Automatic External Defibrillator(AED) in 1997 makes it possible to treat these patients with sudden cardiac arrest outside of hospitalfor general public without any medical knowledge in time and greatly improve survival rate of patient with malignant arrhythmias. Meanwhile, the implement of remote real-time monitoring for patients with arrhythmia as soon as possible will effectively reduce deaths rate of SCD outside the hospital. Therefore, the automatic analysis for abnormal ECG waveform and rapid detection for malignant arrhythmia are the focus and difficulty for not only AED with automatic perception and defibrillation but slso remote ECG monitoring system with real-time monitoring and early warning for high-risk patients.The main work is to focus on remote real-time monitoring for patients with arrhythmia and rapid defibrillation for malignant arrhythmia in this paper, the research are carried out mainly on the following three aspects.Firstly, it describes the generation principle of ECG, the characteristics of ECG complex and the standard database in the world, and then design a band-pass digital filter to remove all kinds of interference and noise in ECG acquisition process.Secordly, it describes the principle and shortcoming of the traditional differential threshold method to detect QRS complex in detail, and then puts forward a real-time QRS complex detection algorithm, which also was designed in the MATLAB and transplanted in STM32 embedded monitoring system, based on differential threshold method, the summarized sensitivity and accuracy of the proposed method based on MIT-BIH database are 99.69% and 99.32%, respectively, which shows a relative performance advantage on the previous algorithms. On the basis of QRS wave detection, 13 kinds of non-fatal arrhythmias in the clinical were detection using rough sets theory.Thirdly, it puts forward four time-frequency analysis detection algorithm based on the shadow and slope variability theory, and then also design and implement in the MATLAB; meanwhile, the standard test database with about 10 million pieces of ECG records was used to evaluate the performance of the four proposed algorithms.the results show that the proposed compositive algorithms would potentially both satisfy requirements by the AHA rules on the arrhythmias detection for AEDs, and show a higher performance and lower computation requirments comparing with the previous HILB algorithm..As a conclusion, the results of this paper are the core technology for the early warning of ECG monitoring and automatic perception and rapid defibrillator of AED. The relevant research will have wide application prospect for automatic analysis and rapid recognition of complex arrhythmias, especially have scientific value and social significance for development of ECG monitoring system and automatic defibrillator device with our own intellectual property rights.